[photorized]

What I always found curious about Twitter, despite its massive concurrency requirements, is how low quality and short-lived the content is. Outside of charting historical trends, there's very little value in mining historical Twitter data - it's only the real-time part that matters. Theoretically, they coould just delete all content older than x days - and I doubt that many would notice.

1bln is funding is unreasonable, both from a technology standpoint (Twitter does go down, btw) and from a "let's figure out a way to monetize this" standpoint.